A Basic Experiment on the Production and Identification of ETG Modes
Wei Xiao,В. А. Соколов,Amiya K. Sen
2007-01-01
Abstract:One of the strongest candidates for the anomalous electron energy transport is believed to be electron temperature gradient (ETG) mode [1, 2]. However, the high frequency (few MHz) and short wave length ( 1 << ⊥ e k ρ ) make the direct observation of ETG modes difficult in experiments. Using a DC bias heating scheme of the core plasma, we are able to produce the drive parameter Te L / 1 above the critical value in Columbia Linear Machine (CLM). A high frequency mode at ~2.3MHz has been observed. Its azimuthal wave number (m ~ 15) and parallel wave number ( 1 // 01 . 0 − ≈ cm k ) has been measured. These values are consistent with the results of a simple kinetic dispersion relation on appropriate 0 0 B E v v × Doppler shift. The scaling of ETG fluctuation level versus Te L , as well as the radial structure of the mode is reported. When the ion transport is strongly suppressed in tokamaks, the particle and energy loss rates for electrons exceed neoclassical transport estimates by one and two orders of magnitude, respectively. The most plausible physics scenario for this anomalous electron transport seems to be based on ETG instabilities. Extensive theoretical and simulation work clearly establish its dynamic behavior, both linear and nonlinear [1-6]. However, the simulation results of the transport consequences remain controversial. Moreover, due to the high frequency and short wave length of ETG modes, no direct experimental evidence of its existence in tokamaks has been found till today [7, 8]. All of the above motivate a basic experiment on the production and identification of ETG modes with a wide range of parameter variation in CLM. CLM is a cylindrical steady-state linear machine that produces a quiescent, magnetically confined, collisionless plasma. The typical plasma parameter in CLM is: 3 9 8 10 5 10 5 ~ − × − × cm n , T B 1 . 0 ≈ , eV Te 15 5 − ≈ , and eV Ti 25 3− ≈ . In order to measure the ETG modes, a Langmuir probe with the good frequency response up to 10 MHz and high spatial resolution (~mm) is specially designed. This frequency response is obtained through minimizing the input capacitance by putting the resistor (SMD 100kΩ) close to the probe tip. Because the magnetic field in CLM is uniform, the Electron Cyclotron Heating (ECH) method doesn’t work. The electron temperature gradient profile is obtained by the discharge current adjustment, DC bias acceleration and thermalization. A positively biased circular tungsten mesh of diameter 1.2cm is placed at the center of the plasma cross-section. The range of the bias voltage used is from +5V to +20V, which accelerates the electrons. The medium neutral pressure in the transition region guarantees that the electrons being accelerated are thermalized to the Maxwellian distribution. And the ions are insensitive to this kind of heating method. The temperature profiles as well as density profiles are shown in Fig. 1. It can be seen that the sharp electron temperature gradient is obtained while ion temperature profile is almost flat. Moreover, the density profile has no gradient within the region of the electron temperature gradient. So the ETG instability in this case is driven by Te L , not ) ln / ln ( n d T d e e = η . The power spectra with different Te L are shown in Fig. 2. It’s clear that the potential fluctuation level increases when Te L decreases, as we expect. For ETG mode, the azimuthal wave number is much larger than the parallel wave number. When we measure the azimuthal wave number through the two very close probe tips (twin probes), the phase difference coming from the z-direction can be ignored. So we fix the radial position r and rotate the twin probes to find the maximum phase shift max φ . Because the maximum phase shift calculated from the cross power spectra may include several π 2 , we rotate the probes continuously and confirm that max φ is real phase shift. If the distance between the two tips of the twin probes is l, the azimuthal wave number can be calculated as ) 360 /( 2 max l r m φ π ⋅ ⋅ = . In our case, o 130 max ≈ φ , cm r 9 . 1 ≈ and cm l 28 . 0 ≈ , so m can be estimated as 15 ~ , which corresponds to 05 . 0 ≈ ⊥ e k ρ . From the phase measurement we can also see that there are about 3 modes with different m hidden inside the peak (|m|=14, 15, 16) in the power spectrum. The minus sign of the phase shift corresponds to the electron diamagnetic direction. The parallel wave number is very difficult to measure because it is mixed up with the much larger azimuthal wave number. We make a very careful probe position calibration by maximizing the plasma conductance between the two probe tips in the different zdirections ( cm Lz 28 ≈ ) and find the parallel wave length is about 4 times of CLM machine length. From this we can estimate that 1 // 01 . 0 − ≈ cm k , which is much smaller